Claude Desktop App
Late yesterday, Anthropic announced messaging support for Claude Code, allowing users to connect to a Claude Code session running on a Mac from a mobile device using Telegram and Discord bots. I spent a few hours playing with it last night, and despite being released as a research preview, the messaging integration is already very capable, but a little fiddly to set up.
Anthropic are out with yet another update to Claude AI: the company’s Claude Code and Cowork tools can now remotely control your Mac on your behalf.
[…]
The capability pairs with Dispatch (released last week) which lets you assign Claude tasks from your iPhone and return to finished work on your desktop. In the YouTube video embedded below, Anthropic’s demo shows a user asking Claude to export a pitch deck as a PDF and attach it to a meeting invite, all while the user is away from their Mac.
[…]
The new feature is essentially Anthropic’s version of OpenClaw, the open-source AI agent that went viral earlier this year.
The Claude Mac client itself remains a lazy Electron clunker. If Claude Code is so good I don’t get why they don’t prove it by using it to make an even halfway decent native Mac app.
Anthropic has released a redesigned Claude Code experience for its Claude desktop app, bringing in a new sidebar for managing multiple sessions, a drag-and-drop layout for arranging the workspace, and more.
[…]
Anthropic has also dropped more of the developer workflow into the app itself. There’s now an integrated terminal for running tests and builds, an in-app file editor for spot edits, a rebuilt diff viewer aimed at large changesets, and an expanded preview pane that handles HTML files and PDFs alongside local app servers. Each pane is also drag-and-drop friendly, so the layout can be arranged to suit.
[…]
In related news, Anthropic also announced Routines – a new way to set up Claude Code automations that run without an active session. A routine bundles a prompt, a repo, and any relevant connectors into a single configuration that can run on a schedule, fire from an API call, or trigger off a GitHub event such as a new pull request.
I strongly suspect Claude’s Mac app is written by Claude.
That’s not a compliment.
[…]
There’s its general everyday bugginess – it frequently resets the scroll position of conversations to some arbitrary point miles back in time, for example. Or just abruptly removes focus from the text field while you’re in the middle of typing (doesn’t move it anywhere else, just defocuses). It smells, in a nutshell.
But the “vibe coding” stench really wafts in when you consider that [cynically] their most important user flow – the upsell – doesn’t even work.
Previously:
- Claude at Apple
- Perplexity Personal Computer
- Codex for Almost Everything
- Gemini App for Mac
- Greg Knauss Is Losing Himself
- OpenClaw (Formerly Moltbot)
7 Comments RSS · Twitter · Mastodon
Still struggling to understand what I’d entrust agentic software to do on my Mac that a macro cannot do more reliably and for free.
"that a macro cannot do more reliably and for free"
Where's the fun in deterministic computing? Are you really alive if you don't risk rm -r / every time your computer does something?
I think I’ll take my standard approach and wait for someone else’s system to get destroyed with a zero day before I put this on anything that matters, like five years from now.
Go nuts, guys. Find out how all this goes horribly wrong for me so I don’t have to.
I use Codex and not Claude Code but Codex is also an Electron app. It's visible that the app is an Electron one but the app is not especially clunky.
I'm wary about letting anything do remote-control my main computer, too.
@Billyok: Codex/Clause Code are not about doing macros. Codex creates code in my style and in good quality. Instead of a more iterative approach to coding I now need to do way more thinking before implementation.
That experience Wade Tregaskis had is basically the experience of all modern software now. Perhaps a bit worse than the average, but this kind of shit is just everywhere.
There's relatively few people who know what makes a good Mac app.
None work at these startups. Almost none work at Apple.
As much as I want to dump on Electron App #78236 for sucking on the Mac, it's not like SwiftUI App #8214 or Catalyst App #8261 are any better on the average.
AI can do anything… except make good Mac apps.
Since Apple basically can't make good Mac apps either, not only will the AI models never train up, an entire generation of users are being trained to accept garbage apps.
You'd think grizzled AppKit and UIKit veterans with decades on the platforms would be a fought over. But the hiring process and management at these companies is a retaining wall for permanent Enshittification.
